AI RESEARCH ANALYSIS

KD-MARL: Resource-Aware Knowledge Distillation in Multi-Agent Reinforcement Learning

Real-world deployment of multi-agent reinforcement learning (MARL) systems is fundamentally constrained by limited compute, memory, and inference time. While expert policies achieve high performance, they rely on costly decision cycles and large-scale models that are impractical for edge devices or embedded platforms. Knowledge distillation (KD) offers a promising path toward resource-aware execution, but existing KD methods in MARL focus narrowly on action imitation, often neglecting coordination structure and assuming uniform agent capabilities. We propose resource-aware Knowledge Distillation for Multi-Agent Reinforcement Learning (KD-MARL), a two-stage framework that transfers coordinated behavior from a centralized expert to lightweight, decentralized student agents. The student policies are trained without critic, relying instead on distilled advantage signals and structured policy supervision to preserve coordination under heterogeneous and limited observations. Our approach transfers both action-level behavior and structural coordination patterns from expert policies with supporting heterogeneous student architectures, allowing each agent's model capacity to match its observation complexity, which is crucial for efficient execution under partial or limited observability along with limited onboard resources. Extensive experiments on SMAC and MPE benchmarks demonstrate that KD-MARL achieves high performance retention while substantially reducing computational cost. Extensive experiments across standard multi-agent benchmarks show that KD-MARL retains over 90% of expert performance while reducing computational cost by up to 28.6× FLOPs. The proposed approach achieves expert-level coordination and can be preserved through structured distillation, enabling practical MARL deployment across resource-constrained onboard platforms.